Thermal Conductivity of Liquid/CNT Core-Shell Nanocomposites
Yutaka Yamada,1,a) Alexandros Askounis,2 Tatsuya Ikuta,2,3 Koji Takahashi,2,3,4 Yasuyuki Takata,2,4,5 and Khellil Sefiane,6,7
1Graduate School of Natural Science and Technology, Okayama University, Okayama 700-8530, Japan
2International Institute for Carbon-Neutral Energy Research (WPI-I2CNER), Kyushu University, Fukuoka 819-0395,
Japan
3Department of Aeronautics and Astronautics, Graduate School of Engineering, Kyushu University, Fukuoka 819-0395,
Japan
4Japan Science and Technology Agency (JST), CREST, Kyushu University, Fukuoka 819-0395, Japan
5Department of Mechanical Engineering, Graduate School of Engineering, Kyushu University, Fukuoka 819-0395, Japan
6Institute for Materials and Processes, School of Engineering, The University of Edinburgh, King’s Buildings, Robert Stevenson
Road, Edinburgh EH9 3FB, UK
7Tianjin Key Lab of Refrigeration Technology, Tianjin University of Commerce, Tianjin City, 300134, PR China
ABSTRACT
Hollow carbon nanotubes (CNTs) were impregnated with an ionic liquid, resulting in a composite core-
shell nanostructure. Liquid infusion was verified by transmission electron microscopy and rigorous
observations unveiled that the nanocomposite is stable, i.e. liquid did not evaporate owing to its low
vapor pressure. A series of individual nanostructures were attached on T-type heat sensors and their
thermal behavior was evaluated. The liquid core was found to reduce the thermal conductivity of the
base structure, CNT, from ca. 28 W/mK to ca. 15 W/mK. These findings could contribute to a better
understanding of nanoscale thermal science and potentially to applications such as nanodevices thermal
management and thermoelectric devices.
I. INTRODUCTION
Carbon nanotubes (CNTs) are typical one-dimensional building materials resembling individual or
multiple graphene sheets rolled up into cylinders1,2. Due to their size and structure, CNTs exhibit unique
properties such as high mechanical strength3, chemical inertness4 and high electrical conductivity5,6. The
thermal properties of individual CNTs, in particular, have attracted considerable scientific attention as
they exhibit extremely high thermal conductivity7-9, with the effect of CNT length10, diameter11,12, defects __________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
and wall structure13,14 also carefully considered. Potentially, CNTs could be employed in high heat flux
applications like nanodevices thermal management15 and/or low heat transport applications like
thermoelectric devices16.
Heat conduction occurs mainly via the CNTs wall rendering the inner, hollow space useless. This space
may be utilized for enhancing or reducing heat conduction by the insertion of a second material which
leads to a core-shell nanocomposite. This is a promising approach as allows the tailoring of the
nanocomposite properties to the application needs without the introduction of defects which may result
in degradation of the nanocomposite. Although, CNTs infusion with a liquid core has been reported
experimentally17,18; little is known about the effect of this infusion on the nanocomposite thermal
behavior, with even theoretical works rather limited19. Previous experimental studies have mainly
focused on a heterostructures which have solid core and solid shell, and the thermal and thermoelectric
performance of core-shell Te−Bi20,21 and Si-Ge22,23 nanowires.
In this paper, we develop a novel core-shell nanocomposite consisting of a liquid core fused inside a
CNT shell structure, confirmed by transmission electron microscopy (TEM). The nanocomposite is then
attached on a nanofilm heat sensor and its thermal conductivity is measured and compared to that of the
base material (CNT). This is the first, to the best of our knowledge, experimental measurement of the
thermal conductivity of liquid filled CNTs.
II. EXPERIMENTAL
The heat nanosensors consisted of a platinum nanofilm suspended on a silicon oxide surface were
fabricated using typical microelectromechanical systems (MEMS) processes. The exact fabrication
procedure is described in detail elsewhere9,24. A scanning electron microscope (SEM; Versa 3D, FEI
Co., Hillsboro, OR, USA) allowed the characterisation of the sensors, with typical 9 µm length, 500 nm
width and 40 nm thickness. In the present study, open-ended, hollow, multi-walled CNTs (US __________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
Nanomaterials; Houston, TX, USA) were selected due to the difficulty of manipulating single-walled
CNTs. A small amount of these CNTs was dispersed in ethanol and gently dropped on a transmission
electron microscopy (TEM) copper microgrid (Ouken Shoji Inc., Japan). Grids were left to dry for 12
hours. Individual CNTs were characterized by high-resolution TEM (JEM-3200FSK; JEOL Ltd., Japan)
operating at an accelerating voltage of 300 kV. A series of CNTs have been measured and Figure 1(a)
shows the micrograph of a typical CNT measured in this study, with ca. 83 and 45 nm outer and inner
diameter, respectively, and ca. 6 m length. Subsequently, the CNTs were impregnated with the ionic
liquid (IL) 1-Butyl-3-methylimidazolium hexafluorophosphate (Kanto Chemical Co., Tokyo, Japan) by
depositing a droplet of the liquid on the microgrid with the CNTs and it spun at 6000 rpm for 1 minute
to remove residual liquid after waited 10 seconds to introduce the liquid into CNTs. This ionic liquid
(IL) was chosen due to: a) its extremely low vapor pressure; no evaporation under electron microscopy
vacuum25 and b) the surface tension of this IL was reported to be ca. 44 mN/m26, which is well below the
surface tension cut-off value reported to be in the range of 100 - 200 mN/m for CNT wetting and
filling17,18,27. TEM observation of a series of CNTs verified the successful impregnation of CNTs with
IL; a micrograph of a characteristic IL filled CNT is presented in Figure 1(b), with ca. 85 and 46 nm as
an outer and inner diameter, respectively. Notably, a number of gas bubbles were observed within the
CNT, as pointed in Figure 1(b), potentially due to slug flow during the capillary filling, in accordance
with the literature28,29 or dissolved gases which emerged in high vacuum condition. The bubbles are not
in contact with the CNT walls due to the high viscosity of IL. Bubble volume fraction was estimated to
be ca. 5 % and is therefore not considered.
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
Figure 1 (a) TEM micrograph depicting part of an empty CNT. (b) TEM micrograph depicting a typical IL-CNT core-
shell nanostructure.
With the help of a manipulation in SEM, individual nanostructures were weakly bonded on the nanotip
of a metallic needle via focused electron beam induced deposition. Subsequently, the nanostructures
were attached first on the nanosensor (NS) and then to the heat sink (HS), as shown in Figure 2 for an
IL/CNT nanocomposite. A constant current was supplied to the NS which acts simultaneously as a
heater and a thermometer. Thus, we are capable of maintaining the whole system at a constant
temperature T0.
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
Figure 2 SEM micrograph depicting the T-type heat nanosensor with an attached IL/CNT core-shell
nanostructure.
III. RESULTS AND DISCUSSION
Our system allows the direct measurement of the volumetric heat generation rate, , which is
given by , where , , and are the heating current, voltage at the NS, width and
thickness of the nanofilm (NS in Figure 2), respectively. Since our measurements are conducted under
the high vacuum conditions of an SEM and the temperature rise was small, both radiation and
convection thermal transport are negligible.9 The thermal conductivity of each nanostructure is thus
calculated using the following formula:9
(1)
where each quantity is represented in Figure 2. and are the nanofilm and CNT cross-sectional
area, is the length of the CNT between NS and HS, is the total length of the NS, and are the
lengths between the CNT junction and the edge of the NS and is the nanofilm (NS) thermal
conductivity. is obtained from , where is the nanofilm electrical resistance
measured at 0oC, is the electrical resistance change during heating and the resistance-
temperature coefficient of the nanofilm as measured during the calibration of the sensor. Figure 3 shows
the thermal conductivity as a function of temperature for the shell (open symbols) and the core-shell
(closed symbols) nanostructures, respectively. In this case the thermal contact resistance, which has been
reported to be small compare to the CNT thermal resistance, was kept to a minimum by minimizing the
shell-sensor junction and depositing an extra layer of amorphous carbon via SEM7,9,30. Hence, the
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
thermal conductivities reported here correspond to the lower bound of the actual intrinsic thermal
conductivities of the nanostructures. The shell nanostructure (CNT) exhibits a λf ≈ 28 W/mK (open
symbols) which is considerably lower than previous reports for similar sized CNTs11,13 but still in line
with the literature for this type of CNTs31,32. In fact, the walls of these CNTs consist of graphene layers
rolled up into cones and stacked one inside the other giving rise to the cup-stacked wall structure33, as
shown in Figure 4. Each graphene layer is inclined a few degrees relative to the longitudinal tube axis.
The effect of this wall structure on the thermal transport of the CNT is discussed elsewhere32. Defects
and interlayer covalent bonding are expected to lower the thermal conductivity, however high-resolution
TEM showed a minimal amount.
Figure 3 Thermal conductivity, , as function of temperature for the shell (open symbols) and the core-shell (closed symbols) nanostructure.
It is readily apparent from Figure 3 (closed symbols) that the liquid core lowers the thermal conductivity
of the nanocomposite from ca. 28 W/mK to ca. 15 W/mK. We shall attempt to address this decrease by
combining conventional 1-D heat conduction in composite materials and phonon heat conduction
mechanism arguments. Heat is conducted following one-dimensional Fourier’s law , with __________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
the heat flow in the axial z direction and the temperature gradient34. This approach is
fundamental to the accurate determination of thermal transport in nanostructures and shows the need for
a precise definition of the CNT cross-sectional area, which remains ambiguous in most experimental
studies due to limitations in measuring the inner wall diameter35. The heat conduction path should, thus,
encounter a thermal resistance in the composite equivalent to34:
(2)
where , and correspond to the thermal resistance of the composite, the liquid core (IL)
and the shell (CNT), respectively. Generally, the thermal conductivity and thermal resistance are related
by: , with heat conduction length , thermal conductivity and surface of conduction ,
which substituted in Eq. (2) yields the effective thermal conductivity of our liquid-core composite:
(3)
where, the thermal conductivity of the core λcore is assumed to be 0.2 W/mK36,37. We used this bulk value
due to a lack of more detailed data in the literature for this particular IL, which should act as the
foundation for our comparison. Additionally, no particular confinement effect is expected, since the
tubular area in our CNT is considerably larger than that of a thin CNT, where the available space in the
tube and the hydrogen bond length become comparable38,39. The thermal conductivity of the shell λshell
was measured to be 28 W/mK and is the total conducting surface of the composite expressed as:
, where Ashell and Acore are the cross sectional area of CNT wall and core liquid,
respectively. Eq. 3 yields an effective thermal conductivity for the composite λeff ≈ 19.9 W/mK. The
lower λeff value can be obtained when λcore is lower than λshell, and this result shows that the thermal
resistance of the sample increased after liquid insertion. Hence, we may conclude that the liquid core is
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
acting as a thermal resistance. To further support this claim, we calculate λshell, using the measured λeff of
the composite, to be approx. 21 W/mK from Eq. (3). This result is smaller than the measured λshell, which
is further evidence of the infused liquid acting as a thermal resistance. Moreover, these calculations
show the importance of accurately defining the cross-sectional area for heat conduction, which is further
supported when considering phonon heat conduction in nanostructures. Nonetheless, these values are
close but not equal to the measured one.
Figure 4 Typical TEM micrograph depicting the inclined wall structure of the CNTs in this work.
This discrepancy shows the limitation of the conventional heat conduction mechanism for
nanomaterials10, 40-42, leading us to further consider the limiting factors of phonon-phonon and phonon
boundary scattering, as phonons are the main energy carrier regardless of diameter (for a comprehensive
review see Ref. 35). In fact, Chang et al., reported a similar deviation of the thermal conductivity of
carbon and boron nanotubes from Fourier’s law due to isotopic deviations10. Nonetheless, Fourier’s law
remains a useful tool to approximate the effective thermal conductivity of nanomaterials. For example,
Thomas et al. predicted, based on Fourier’s law19, a decrease in the effective thermal conductivity of a
single-walled carbon nanotube when a liquid core was inserted. In particular, they estimated that the
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
vibrational frequency of the water atoms coincides with the low vibrational frequencies of the acoustic
phonons. Therefore, interactions between the two should lead to phonon scattering and in turn to a
decrease in the thermal conductivity of the composite19. Our experimental results are complementary to
this theoretical work, therefore we expect a similar thermal transport mechanism to be at work, with the
additional effect of the unique wall structure. Fig. 4 depicts this unique structure and the white line
highlight the inclination of the graphene layers in relation to the tube axis. This inclination has a
significant effect on the heat transport of the core structure as the phonons follow the ballistic regime
within each graphene layer and the diffusive across the tube length, as we have discussed in detail
elsewhere32. The composite thermal transport should also be affected by this unique wall structure.
Specifically, more graphene edges are in contact with the liquid which should amount to a higher
amount of phonon scattering giving rising to the observed discrepancy. This argument merits further
exploration in the future.
IV. SUMMARY AND CONCLUSIONS
To summarize, we have successfully prepared a liquid-core nanocomposite. Using a thin film heat
nanosensor we evaluated the thermal conductivity of the nanocomposite and its base structure a CNT. A
decrease in the thermal conductivity of the nanocomposite was found and it was lower than the result
predicted by bulk scale theory, due to phonon interaction with the liquid molecules. This is the first, to
the best of our knowledge, experimental evidence of a liquid/CNT core-shell nanocomposite and its
thermal behavior assessment. We believe that these results contribute to a better understanding of
nanoscale thermal transport. Additionally, we provided experimental evidence and quantification of heat
transfer at the solid-liquid interface of nanocomposites. Potentially, our results could pave the way for
further research into nanoscale phase change phenomena, chemical reactions, fluid flows and
thermoelectrics in CNTs.
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
ACKNOWLEDGEMENTS
We acknowledge the Japan Society for the Promotion of Science (JSPS) for the Postdoctoral Fellowship
for North American and European Researchers. This work was financially supported in part by the Core
Research for Evolutional Science and Technology project of Japan Science and Technology Agency
(JST-CREST). We are also grateful to the research laboratory for High Voltage Electron Microscopy at
Kyushu University, for use of the TEM facilities.
Author information*Corresponding AuthorYutaka YamadaTel.: +81-86-251-8046; Fax: +81-86-251-8266;E-mail: [email protected]
NotesThe authors declare no competing financial interest.
References1 S. Iijima, Nature 354, 56 (1991).2 S. Iijima, and T, Ichihashi, Nature 363, 603 (1993).3 B. I. Yakobson, C. J. Brabec, and J. Bernholc, Phys. Rev. Lett. 76, 2511 (1996).4 H. Dai, Acc. Chem. Res. 35, 1035 (2002).5 J. W. G. Wildöer, L. C. Venema, A. G. Rinzler, R. E. Smalley, and C. Dekker, Nature 391, 59 (1998).6 C. Dekker, Phys. Today 52, 22 (1999).7 P. Kim, L. Shi, A. Majumdar, and P. L. McEuen, Phys. Rev. Lett. 87, 215502 (2001).8 C. Yu, L. Shi, Z. Yao, D. Li, and A. Majumdar, Nano Lett. 5, 1842 (2005).9 M. Fujii, X. Zhang, H. Xie, H. Ago, K. Takahashi, T. Ikuta, H. Abe, and T. Shimizu, Phys. Rev. Lett. 95,
065502 (2005).10 C. W. Chang, D. Okawa, H. Garcia, A. Majumdar, and A. Zettl, Phys. Rev. Lett. 101, 075903 (2008).11 J. Yang, S. Waltermire, Y. Chen, A. A. Zinn, T. T. Xu, and D. Li, Appl. Phys. Lett. 96, 023109 (2010).12 M. T. Pettes and L. Shi, Adv. Funct. Mater. 19, 3918 (2009).13 H. Hayashi, T. Ikuta, T. Nishiyama, and K. Takahashi, J. Appl. Phys. 113, 014301 (2013).14 H. Hayashi, K. Takahashi, T. Ikuta, T. Nishiyama, Y. Takata, and X. Zhang, Appl. Phys. Lett. 104, 113112
(2014).15 E. Pop, Nano Res. 3, 147 (2010).16 C. Meng, C. Liu, and S. Fan, Adv. Mater. 22, 535 (2010)17 E. Dujardin, T. W. Ebbesen, H. Hiura, and K. Tanigaki, Science 265, 1850 (1994).18 B. M. Kim, S. Sinha, and H. H. Bau, Nano Lett. 4, 2203 (2004).19 J. A. Thomas, R. M. Iutzi, and A. J. H. McGaughey, Phys. Rev. B 81, 045413 (2010).20 G. Zhang, W. Wang, and X. Li, Adv. Mater. 20, 3654 (2008).21 J. Kang, J. W. Roh, W. Shim, J. Ham, J.-S. Noh, and W. Lee, Adv. Mater. 23, 3414 (2011).22 M. Hu, X. Zhang, K. P. Giapis, and D. Poulikakos, Phys. Rev. B 84, 085442 (2011).__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]
23 M. C. Wingert, Z. C. Y. Chen, E. Dechaumphai, J. Moon, J.-H. Kim, J. Xiang, and R. Chen, Nano Lett. 11, 5507 (2011).
24 X. Zhang, H. Xie, M. Fujii, H. Ago, K. Takahashi, T. Ikuta, H. Abe, and T. Shimizu, T. Appl. Phys. Lett. 86, 171912 (2005).
25 T. Welton, Chem. Rev. 99, 2071 (1999).26 M. G. Freire, P. J. Carvalho, A. M. Fernandes, I. M. Marrucho, A. J. Queimada, and J. A. P. Coutinho, J.
Colloid Interface Sci. 314, 621 (2007).27 T. W. Ebbesen, J. Phys. Chem. Solids 57, 951 (1996).28 S. Chen, G. Wu, M. Sha, and S. Huang, J. Am. Chem. Soc. 129, 2416 (2007).29 G. Brown, S. R. Bailey, M. Novotny, R. Carter, E. Flahaut, K. S. Coleman, J. L. Hutchison, M. L. H. Green,
and L. Sloan, Appl. Phys. A 76, 457 (2003).30 L. Shi, D. Li, C. Yu, W. Jang, D. Kim, Z. Yao, P. Kim, and A. Majumdar, J. Heat Transfer 125, 881 (2003).31 K. Takahashi, Y. Ito, T. Ikuta, X. Zhang, and M. Fujii, Physica B: Condensed Matter 404, 2431 (2009).32 A. Askounis, Y. Yamada, T. Ikuta, K. Takahashi, Y. Takata, and K. Sefiane, AIP Advances 6, 115119 (2016)33 M. Endo, Y. A. Kim, T. Hayashi, Y. Fukai, K. Oshida, M. Terrones, T. Yanagisawa, S. Higaki, and M. S.
Dresselhaus, Appl. Phys. Lett. 80, 1267 (2002).34 T. L. Bergman, F. P. Incropera, D. P. DeWitt, and A. S. Lavine, Fundamentals of Heat and Mass Transfer.
(Wiley, 2011).35 A. M. Marconnet, M. A. Panzer, and K. E. Goodson, Rev. Mod. Phys. 85, 1295 (2013).36 M. E. V. Valkenburg, R. L. Vaughn, M. Williams, and J. S. Wilkes, Thermochim. Acta 425, 181 (2005).37 A. P. Fröba, M. H. Rausch, K. Krzeminski, D. Assenbaum, P. Wasserscheid, and A. Leipertz, Int. J.
Thermophys. 31, 2059 (2010).38 K. Dong, G. Zhau, X. Liu, X. Yao, S. Zhang, and A. Lyubartsev, J. Phys. Chem. C 113, 10013 (2009).39 R. Singh, J. Monk, and F. R. Hung, J. Phys. Chem. C 114, 15478 (2010).40 N. Mingo and D. A. Broido, Nano Lett. 5, 1221 (2005).41 A. L. Moore, and L. Shi, Mater. Today, 17, 163 (2014).42 T. Mai, and O. Narayan, Phys. Rev. E, 73, 061202 (2006).
__________________________________________________________________________________
a) Author to whom correspondence should be addressed. Electronic mail: [email protected]